@goodandready/dsh-moa 0.2.41 → 0.2.43
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +7 -0
- package/README.ru.md +7 -0
- package/README.zh.md +11 -0
- package/lib/client.js +421 -107
- package/lib/file-workspace.js +40 -43
- package/lib/history.js +8 -1
- package/lib/index.js +55 -4
- package/lib/live-canvas.js +39 -6
- package/lib/moa-budget.js +28 -19
- package/lib/moa-candidates.js +3 -0
- package/lib/moa-multi-judge.js +2 -1
- package/lib/moa-prompts.js +5 -1
- package/lib/moa-runner.js +48 -46
- package/lib/moa-stream.js +59 -1
- package/lib/moa-test-gate.js +11 -0
- package/lib/routes.js +10 -3
- package/lib/updater.js +23 -0
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -116,6 +116,13 @@ The flag form is equivalent:
|
|
|
116
116
|
/moa --preset=deep-reasoning solve this math problem step by step
|
|
117
117
|
```
|
|
118
118
|
|
|
119
|
+
Or promote a candidate solution directly (by candidate index or runId + candidate index):
|
|
120
|
+
|
|
121
|
+
```text
|
|
122
|
+
/moa promote 2
|
|
123
|
+
/moa promote <runId> 2
|
|
124
|
+
```
|
|
125
|
+
|
|
119
126
|
### 2. Adaptive Questionnaire Gate
|
|
120
127
|
When prompts are open-ended or lack architectural specifications (e.g. *"build a calculator app"*), advisor models detect ambiguities and formulate focused clarifying questions (e.g., UI style, persistence backend, framework choice) before generating code.
|
|
121
128
|
|
package/README.ru.md
CHANGED
|
@@ -116,6 +116,13 @@ graph TD
|
|
|
116
116
|
/moa --preset=deep-reasoning реши эту математическую задачу по шагам
|
|
117
117
|
```
|
|
118
118
|
|
|
119
|
+
Или прямое продвижение решения кандидата (по индексу кандидата или runId + индексу):
|
|
120
|
+
|
|
121
|
+
```text
|
|
122
|
+
/moa promote 2
|
|
123
|
+
/moa promote <runId> 2
|
|
124
|
+
```
|
|
125
|
+
|
|
119
126
|
### 2. Адаптивный гейт уточнения требований
|
|
120
127
|
Когда запрос сформулирован слишком обобщенно (например, *"сделай калькулятор"*), советники определяют недостающие архитектурные требования и формулируют целевые вопросы (стиль интерфейса, сохранение состояния, стек технологий) до генерации кода.
|
|
121
128
|
|
package/README.zh.md
CHANGED
|
@@ -116,6 +116,13 @@ graph TD
|
|
|
116
116
|
/moa --preset=deep-reasoning 逐步求解这道数学题
|
|
117
117
|
```
|
|
118
118
|
|
|
119
|
+
或直接提升候选者方案（支持指定 runId 与候选序号）：
|
|
120
|
+
|
|
121
|
+
```text
|
|
122
|
+
/moa promote 2
|
|
123
|
+
/moa promote <runId> 2
|
|
124
|
+
```
|
|
125
|
+
|
|
119
126
|
### 2. 自适应澄清问卷
|
|
120
127
|
当需求过于抽象（例如 *"制作一个计算器"*）时，系统会在生成代码前主动询问 UI 风格、数据持久化方式或框架偏好。
|
|
121
128
|
|
|
@@ -128,6 +135,7 @@ graph TD
|
|
|
128
135
|
* 各提案模型在 `.moa/candidate-1/`、`.moa/candidate-2/` 等独立目录生成工程代码。
|
|
129
136
|
* 裁判模型对比各版本实现，通过 `WINNER_CANDIDATE_INDEX: N` 指定最优方案。
|
|
130
137
|
* 胜出方案自动提升至项目根目录，临时沙箱随后自动清理。
|
|
138
|
+
* 支持随时通过命令 `/moa promote [runId] <candidateIndex>` 或 UI 界面一键手动提升指定候选者的方案。
|
|
131
139
|
|
|
132
140
|
### 5. 原生设置卡片与预设管理
|
|
133
141
|
在 `设置 → 插件 → Mixture of Agents` 中可视化配置模型：
|
|
@@ -234,6 +242,9 @@ dsh-moa:
|
|
|
234
242
|
| `/dsh-moa/leaderboard` | `GET` | 计算模型胜率排行榜与平均执行成本 |
|
|
235
243
|
| `/dsh-moa/runs/<id>` | `GET` | 按 id 返回单条运行记录 |
|
|
236
244
|
| `/dsh-moa/run` | `POST` | 通过 HTTP 运行完整 MoA 管线（`enabled: false` 时返回 400） |
|
|
245
|
+
| `/dsh-moa/diff` | `GET` | 计算候选运行结果或综合输出的行级别差异（支持按 runId 解析工作区） |
|
|
246
|
+
| `/dsh-moa/promote` | `POST` | 手动将候选工作区文件提升到项目根目录（带时间机器快照） |
|
|
247
|
+
| `/api/dsh-moa/update` | `POST` | 从 npm 一键更新插件，带安全写入源校验 |
|
|
237
248
|
|
|
238
249
|
---
|
|
239
250
|
|